It just doesn't make sense to buy every single EA that comes out. There is expense, not to mention the possibility it is a scam and blows your account (the ol' scam EA double-dip). That's why we must have a set of procedures to follow in deciding whether we should ever expose our capital to these programs. I see them as employees you can use to go out and make you money. You want the same qualities. Someone who doesn't risk your capital in a dangerous way, someone reliable, someone profitable. Remember, YOU are the boss and THEY are your employees and you only want good people working for you.
So what are we looking for OR looking to avoid in an EA???
1. EA Type
We DO NOT want any grid traders, martingale or systems we cannot understand.
We DO want EAs that are suited for the types of markets and trades that we are NOT taking manually. First obvious reason is not to take the same trade twice, but we want to make the most of all our market time and having an EA trade the Asian session while you're sleeping or doing analysis for the European session is effective use of that time. You can focus on your big, manual trades while your EAs work markets you are either are not familiar with, aren't awake to trade or styles such as scalping that you don't trade manually. A good EA portfolio should COMPLEMENT your own personal trading style and add depth to the trading by diversification (as we all know, our manual systems don't work ALL the time). We also want EAs to trade styles that we simply cannot do without being stuck in front of the monitor with your "finger on the trigger" for hours waiting for a two-second opportunity (volatility expansion and news trading for example).
2. Performance
What is a good EA without performance?? What we are looking for from a numbers standpoint are
criteria although they don't all need to be working together to make a good EA. Sometimes just the right amount of one of these will do:
- Positive Reward/Risk Ratio. We are looking for an EA that makes bigger wins than it takes in losses. These are usually NOT scalpers. Forex Growth Bot is an excellent example of this. It's wins are usually in the 100+ pip range and losses usually around 30+ pips. This gives us a positive R/R of over 3/1. So we can win one and lose two and still make money. The best way to analyze this is to take the average win and compare it to the average loss.
- High Win Percentage. This is CRUCIAL for EAs that do not have a positive R/R ratio. It is OK that wins are smaller than losses as long as the EA wins a high percentage to offset the occasional loss.
- Smooth Equity Curve. This is for slightly longer-term analysis of the EA. When an EA makes consistent, predictable trades, you get an equity curve that looks very much like a bending line that goes up (if it goes down smoothly, we are not interested!). When an EA does not trade consistently and occasionally goes through very rough waters and its wins are sometimes big and sometimes small then you will see a very choppy equity curve with many zig-zags. Essentially smooth equity curve = much less fiddling with settings all the time because it's relatively consistent.
- Average Consecutive Loss/Win. Again, this should be taken into account in light of what kind of system it is but should be used in calculating risk. If the average consecutive loss is 5, then be prepared for 5 losses in a row to occur regularly and compensate risk accordingly. I personally usually like a higher average consecutive win over consecutive loss unless the wins are extremely larger than the losses but even at that who wouldn't like better consecutive huge wins?
3. Money Management (Probably most important but not always readily discernible)
- Easy to understand risk settings. You should ALWAYS, ALWAYS, ALWAYS know how much you are risking per trade. PERIOD. Did I mention this was important?
- Adjustable based on either % of equity OR fixed lot. This is standard in good EAs. You know you got a cheap one if there is not a choice of both of these two options.
- Catastrophic trade action. There are times where your trades will fail, AND FAST!! What happens if the stop gets jumped? Can the EA respond? Some EAs have a built-in "circuit breaker" that if something catastrophic happens like getting 50% of your account blown out due to very rare, and unpredictable events like BOJ intervention or say, a tsunami. Many have a circuit breaker for margin levels as well. You ALWAYS want to know the worst case scenario and how your EAs will deal with it. This is not always easy to find out NOR is it common in many EAs.
- Dynamic stop losses and t/p. There are many EAs that adjust stops and take profits to the volatility of the market. This is a very nice feature to have if it works correctly. Always is better to get that stop to entry or better as soon as possible within guidelines of the trading system.
- Huge stop losses that don't make sense. Sometimes EAs will simply use gigantic stop losses to "curve-fit" their EA to historical charts. You know, if you had a big enough stop loss or enough capital not to have to use one (not many people out there in this position), then almost every trade you took, regardless of direction would win at some point. We don't live in that world though (not yet at least!) and a gigantic stop loss that is placed for no reason other than to increase win % is a big no-no. Also can be said for very small s/l unless they fit the system. Remember that if someone wants to make the average loss look very small, they just have to reduce the stop without changing the order size. But when we check the average consecutive number of win/loss AND the overall profitability of the system, we can see if any funny business is going on in this department.
- Adjustable stop and take-profit. DANGER: Adjusting these settings without a very deep knowledge of the system and market you are trading is NOT RECOMMENDED. If it's a good system, then these shouldn't have to be modified ever. BUT, IF, you have used a EA long enough to really feel confident that you could improve the performance a little, then this is where you go. For example, I have adjusted my FapTurbo s/l because I have seen enough of its work to know I can shave a few pips off the s/l to improve the risk/reward ratio without killing or really damaging its win %. We will be getting to settings on an EA by EA basis, so no need to worry about this but everyone should be aware these features exist.
4. Good Vendor Track Record
If someone scams you once, they will probably repackage garbage and try to do it again. This is not always an option if it's a new vendor but something that I check with ALL EAs. It can also work in a positive way, and if you get something profitable from a vendor, then chances are, future products may also be good. I actually prefer to see EAs from people that have already produced profitable ones.
5. Refund Policy
Sometimes and EA may not work for your broker, or maybe it doesn't fit your trading profile. Or maybe it just doesn't make any money. Why should you have to be out your money that you spent on it? When you purchase products through Clickbank or Plimus, you know you can get a refund even with a stinker of a vendor and product. I recommend always using these because they GUARANTEE you get your money back within the period set by the vendor. I have NEVER had a problem getting a refund through either of these two payment processors. That doesn't mean the vendor won't do everything they can to keep your money, like changing your refund request to a support ticket 3-4 times. But in the end, you can get your refund. I also like vendors that give you 60 days to try the product or offer a free trial period. 60 days gives you enough time to run on a demo, and then move to a small, live account and see performance before expiry of refund.
In the next post I will explain my "filtration" process where I go through and choose and test and EA from start to finish and the steps I take to ensure that I at least have a good chance of success with any product I actually decide to purchase.
1Likes
LinkBack URL
About LinkBacks
Rules of engagement
Reply With Quote
